Hello, I’m Duc Nguyen (Duke)
Today I will introduce ML.NET to the .NET community
So, what’s the ML.NET?
- ML.NET is a cross-platform machine learning framework developed by Microsoft
- Open source
- Used in situations where it’s necessary to incorporate machine learning into C# or F# applications in order to avoid switching to other programming languages, such as Python (.NET ecosystem).
- Able can be expanded to function with additional machine learning libraries, like TensorFlow, assisting in utilizing current models and lowering obstacles to ML.NET adoption.
- Can be deployed in various environments, from desktop apps to web services
Alright, first you can get the source code from my repository here
Then, you can see:
– HaNoi-VN_housing_dataset.csv: sample data file used for training
– lib/HousePriceForecast.consumption.cs: Defines the “Predict” method, which takes input data and returns predictions using the trained model
– lib/HousePriceForecast.evaluate.cs: Defines the method for calculating PFI (Permutation Feature Importance) through the “CalculatePFI()” method — determines the level of influence from input data
– lib/HousePriceForecast.mlnet: Binary file containing the trained ML.NET model and related metadata and if there’s a similar project, this model can be reused without retraining
– lib/HousePriceForecast.training.cs: Used to train the ML.NET model and defines data processing steps and algorithm selection
Now, on to the fun part, practice!
You can also refer to the animation image here